Allergies to sulfonamide antibiotics and related drugs

INTRODUCTION — Sulfonamide-containing antibiotics are the second most frequent cause of allergic drug reactions, after the beta-lactams (penicillins and cephalosporins) [1] . The adverse reactions caused by sulfonamides differ significantly from those attributed to beta-lactams, however, and the evaluation and management of sulfonamide reactions is distinct. The pathophysiology of allergic (or hypersensitivity) reactions to sulfonamides is complex and poorly understood.

This topic review describes the most prevalent types of reactions to sulfonamides, the evaluation of patients with reported allergy, and cross-reactivity issues. Management options for patients with past reactions who require similar medications again are also presented. Most of the available literature concerns allergic reactions to antimicrobial sulfonamides, particularly trimethoprim-sulfamethoxazole (TMP-SMX). However, reactions to non-antimicrobial sulfonamides and sulfones will be briefly addressed as well.

This review does not specifically address the treatment of acute drug reactions. These are presented elsewhere, according to specific type of reaction. (See "Stevens-Johnson syndrome and toxic epidermal necrolysis: Management, prognosis, and long-term sequelae" and see "Drug eruptions" and see "Anaphylaxis: Rapid recognition and treatment" and see "Drug fever").

TERMINOLOGY — The imprecise term "sulfa drugs" is most often applied to sulfonamide antimicrobials, although it is variably applied to other medications as well. Widespread use of this term has contributed to ongoing confusion about relevant cross-reactivity. It is preferable to avoid "sulfa" and instead document the specific medication that caused the adverse reaction and the symptoms that were involved.

Sulfonamides — Sulfonamide medications are drugs that contain a sulfonamide moiety (SO2NH2) [2] . There are two distinct groups of sulfonamides that differ in chemical structure as well as clinical use.

Antimicrobial (arylamine) sulfonamides — Antimicrobial sulfonamides contain an arylamine group (an amine group linked to a benzene ring). This moiety is attached to the sulfonamide group (show figure 1). The presence of the arylamine moiety is believed central to the pathogenesis of hypersensitivity reactions, and only the antimicrobial sulfonamides contain it. These drugs also contain a substituted aromatic ring attached to the sulfonamide core (show figure 1). Included in this group of medications are sulfamethoxazole, sulfisoxazole, and sulfadiazine (show figure 2A-2B, upper portions) [2] .

Non-antimicrobial (non-arylamine) sulfonamides — The non-antimicrobial sulfonamides do not contain an arylamine group or a substituted aromatic ring. Members of this group include diuretics, hypoglycemic and antihypertensive agents, as well as celecoxib, sulfasalazine, and sumatriptan (show figure 2B-2B, lower portions) [2] .

Sulfones — Sulfones are a distinct class of medications that are not sulfonamides. Dapsone (diaminodiphenylsulfone) is the only sulfone in common clinical use (show figure 3). The hypersensitivity reactions reported with sulfones are clinically similar to those of sulfonamide antimicrobials, may have similar pathogenesis, and may have some cross-reactivity concerns as well. For these reasons, sulfones are included in this topic review.

Medications sometimes confused with sulfonamides — The following groups of medications are unrelated to sulfonamides: Some medications contain sulfur atoms in their chemical structure but do not contain specific sulfonamide or sulfone moieties. These are not generally thought of as "sulfa" medications and include amoxicillin, captopril, omeprazole, ranitidine, and others [3] . Another group of medications contains sulfate ions, such as albuterol sulfate. The sulfate moieties in these compounds are not a known source of allergic or immunologic issues. Finally, sulfites are chemicals used to preserve certain foods, beverages, and medications. Sulfites are associated with a distinct constellation of symptoms known as sulfite sensitivity, which is primarily mediated via non-immunologic mechanisms. Reactions to these latter medications and substances are reviewed separately. Sulfite sensitivity typically causes respiratory reactions in patients with asthma and is unrelated to sulfonamide hypersensitivity.

Adverse drug reactions — It is also important to clarify terminology regarding adverse drug reactions. An adverse drug reaction may be defined as any undesired and unintended response that occurs at appropriate doses of a drug, given for the therapeutic, diagnostic, or prophylactic benefit of the patient, and occurring within a reasonable time after administration of the drug (excluding therapeutic failure) [4] . Adverse reactions can be further categorized as predictable versus unpredictable. Predictable reactions include "side-effects" and indirect secondary effects. Unpredictable reactions are usually dose-independent, unrelated to the drug's pharmacological actions, and may be influenced by the patient's immunologic responsiveness [4] .

Only unpredictable reactions that are known (or strongly presumed) to be immunologic in nature are classified as allergic (or hypersensitivity) reactions, and these reactions account for a minority of all adverse drug reactions. Approximately 6 percent of inpatient adverse drug reactions were due to allergic reactions in one large series [5] .

EPIDEMIOLOGY — The epidemiology of sulfonamide reactions has changed markedly with the appearance of human immunodeficiency virus (HIV) infection.

Prior to the acquired immune deficiency syndrome (AIDS) epidemic, the prevalence of adverse reactions was relatively low. As an example, in a large, multicenter review published before the epidemic, adverse reactions to TMP-SMX occurred in 8 percent of 1,121 hospitalized patients receiving oral or parenteral forms of this medication [6] . Women were affected twice as often as men.

By comparison, in patients with HIV infection, cutaneous drug reactions to both sulfonamides (and TMP-SMX specifically) and sulfones are markedly more common [7] . Early reports on HIV-infected patients requiring TMP-SMX for Pneumocystis carinii (PCP, since renamed P. jirovecii) prophylaxis found that as many as 34 percent discontinued the drug because of fever and rash [8] . In the setting of both active AIDS and PCP, reactions to TMP-SMX were reported in as many as 50 percent [9] . Several clinical and immunological factors have been hypothesized to enhance the risks of sulfonamide reactions in this population, as discussed below [10,11] . (See "Pathophysiology" below).

Little has been published on changes in the epidemiology of sulfonamide reactions since the advent of highly active anti-retroviral therapy (HAART), although the incidence of these reactions appears to have decreased precipitously. One possibility is that fewer patients are requiring regular prophylaxis with TMP-SMX for P. jirovecii because HAART can preserve lymphocyte counts. It is unclear if other factors are involved.

TYPES AND CLINICAL MANIFESTATIONS OF HYPERSENSITIVITY REACTIONS

Non HIV-infected patients — In patients without HIV infection, adverse reactions to sulfonamides and sulfones are uncommon, as mentioned previously, and only a subset are clearly immunologic in nature. In one large study of 1,121 hospitalized patients, adverse reactions to these medications were most often gastrointestinal (4 percent) or dermatologic (3 percent) [6] . Gastrointestinal symptoms included nausea, vomiting, anorexia, and diarrhea. Dermatologic symptoms included erythema, urticaria, and pruritis. Other reactions, including fever, thrombocytopenia, and renal tubular acidosis were reported in less than 1 percent. No cases of exfoliative skin reactions were noted. The majority of these reactions appeared within the first three days of therapy and nearly all resolved promptly after drug discontinuation.

HIV-infected patients — Individuals with HIV infection commonly develop reactions to these medications and most are immunologic, in contrast to the observations in non-HIV infected patients. The most prominent hypersensitivity reactions observed with sulfonamide antibiotics in this population are morbilliform rash and fever, sometimes accompanied by other organ involvement.

Sulfonamides and sulfones can also cause the spectrum of rare reactions encompassing Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN). These reactions are more common in HIV-infected patients as well. None of these hypersensitivity reactions fit neatly into the classification scheme for immunologic mechanisms established by Gell and Coombs, and are sometimes included in a fifth category of reactions of uncertain mechanism (show table 1).

Fever and morbilliform rash — Sulfonamide hypersensitivity (particularly in the setting of HIV-infection) is characterized by fever and pruritic maculopapular or morbilliform rash, occasionally followed by multi-system organ involvement and inflammation. Severity can range from mild to severe, with fatalities reported.

Symptoms generally develop one to two weeks after the start of administration, with fever usually appearing first, sometimes accompanied by malaise and pharyngitis. Peripheral blood smear may reveal an atypical lymphocytosis or eosinophilia [12] . Organ involvement may be asymptomatic or overt, and can include hepatitis, nephritis, pulmonary infiltrates, and cytopenias. Symptoms generally resolve one to two weeks after discontinuation [2,13] . With re-exposure, symptoms may appear as soon as one to two days.

Other causes of fever and rash in HIV-infected individuals are discussed separately. (See "Fever and rash in HIV-infected patients").

Stevens-Johnson syndrome and toxic epidermal necrolysis — Antimicrobial sulfonamides are the most common medications etiologically associated with Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN, or Lyell syndrome). In contrast, the non-antimicrobial sulfonamides (including furosemide, thiazide diuretics, and sulfonylurea hypoglycemics) do not appear to be associated with excess risk of SJS/TEN [14] .

SJS is characterized by a prodrome of malaise and fever, followed by the rapid onset of erythematous or purpuric macules and plaques and blistering of the mucous membranes. The skin lesions progress to epidermal necrosis and sloughing. TEN, or Lyell's syndrome, is a similar but more severe disorder that involves a greater percentage of the body surface area.

The immunologic mechanisms of SJS/TEN remain enigmatic. These disorders are discussed in detail separately. (See "Stevens-Johnson syndrome and toxic epidermal necrolysis: Clinical manifestations, pathogenesis, and diagnosis").

Erythema multiforme (EM) is a milder cutaneous reaction consisting of target and vesiculobullous lesions involving the mucosal membranes and favoring the extremities (and palms and soles). SJS and severe erythema multiforme (EM) are viewed by many experts, although not all, as separate entities. EM has also been reported in association with sulfonamide antibiotics [15,16] . (See "Drug eruptions").

Other uncommon reactions — Sulfonamide drugs are infrequently associated with other types of immunologic reactions. Type I, or immediate-type hypersensitivity reactions (show table 1) [17,18] . These reactions are mediated primarily by specific IgE antibodies, which trigger sensitized mast cells (and basophils) to degranulate. Histamine, as well as rapidly synthesized vasoactive mediators (such as leukotrienes), are released. For many drugs, including sulfonamides, the relevant drug antigen is not the intact drug itself. Instead, reactive metabolites may become bound in large numbers to serum or cell-surface carrier proteins and thereby create a multivalent immunologic stimulus, which can differ structurally from the parent drug. This fact substantially limits the utility of simple skin testing with the original (unmetabolized) sulfonamide, which does not elicit positive wheal-and-flare responses even in the most convincing cases of sulfonamide anaphylaxis. (See "Allergy to penicillins"). Another type of reaction seen uncommonly with sulfonamide antimicrobials is a Type III, or serum-sickness like reaction (show table 1). These present approximately 10 days to two weeks after initiation of therapy, with fever, rash (which is often urticarial), arthralgia, and lymphadenopathy [19] . (See "Serum sickness and serum sickness-like reactions"). Sulfonamides can cause hemolytic anemia in patients with glucose-6-phosphate dehydrogenase (G6PD) deficiency and are contraindicated in patients known to have this condition. G6PD deficiency is a genetic metabolic-type adverse drug reaction rather than an immunologic reaction, and it is presented elsewhere. (See "Clinical manifestations of glucose-6-phosphate dehydrogenase deficiency").

PATHOPHYSIOLOGY — The pathophysiology of the most prevalent type of sulfonamide hypersensitivity reaction (ie, fever and non-exfoliative rash) has not been conclusively defined and is likely multifactorial [3,10,13] . Metabolism of sulfonamide antimicrobials results in intermediate forms of the drug that are believed to be critical in the pathogenesis of these reactions. It is unclear if the mechanisms described herein are involved in fever and non-exfoliative rash only, or also in SJS/TEN, although cytotoxic mechanisms are likely more prominent in SJS/TEN. (See "Stevens-Johnson syndrome and toxic epidermal necrolysis: Clinical manifestations, pathogenesis, and diagnosis").

Sulfonamide antimicrobials may be metabolized along several pathways. Some of the drug becomes acetylated (via N-acetyl transferase) at the arylamine moiety to form non-toxic metabolites that are excreted in the urine. Alternatively, the arylamine group may undergo N-oxidation (via cytochrome P450) to form reactive metabolites (show figure 3) [10,20] . These reactive metabolites may act as haptens, or small molecules that become covalently linked to large host proteins. These hapten-protein complexes may be recognized as foreign and initiate an immunologic reaction. Other intermediate metabolites may be directly cytotoxic [10,20] .

Factors that slow metabolism, such as slow acetylation and/or glutathione deficiency states (such as HIV infection), may increase exposure to these metabolites and thus enhance the likelihood of hypersensitivity reactions (show figure 4) [11] . (See "Drugs and the liver: Metabolism and mechanisms of injury").

CROSS-REACTIVITY — Studies measuring cross-reactivity among related medications are often confounded by the fact that a clinical history of drug allergy, by itself, appears to identify a subgroup of patients that is at increased risk of reactions to medications in general, including those to distinct and non-crossreactive classes [21] . Patients with reactions to several different types of drugs are sometimes said to have "multiple drug allergy syndrome."

The propensity of certain patients to develop drug reactions was demonstrated in a large, retrospective cohort study performed on a medical database of over 8 million patients over a 12 year period [22] . Individuals who had a documented allergic reaction to a sulfonamide antibiotic did react more commonly to a sulfonamide non-antibiotic (10 percent) than others who tolerated sulfonamide antibiotics in the past (1.6 percent). However, those same individuals with documented sulfonamide antibiotic reactions reacted to the chemically distinct penicillins even more often (14 percent).

This strongly suggests that the reactions observed to the other sulfonamides resulted from a general propensity for immunologic drug reactions, as opposed to specific cross-reactivity between different groups of sulfonamides. The converse was also noted: there was a higher risk of reaction to sulfonamide non-antibiotics in those with a history of reactions to penicillins than those with a history of reactions to sulfonamide antibiotics.

The conformation of the major drug allergens in sulfonamide reactions has not been identified and thus skin testing and serological studies have not been possible. Instead, cross reactivity between sulfonamide medications has to be inferred entirely from clinical studies of cross-sensitivity, which are often confounded, as described previously.

Between sulfonamide antimicrobials and non-antimicrobials — There is minimal evidence of cross-reactivity between sulfonamide antimicrobials and non-antimicrobials (show figure 2A-2B). Most, although not all, patients with a history of reacting to an antimicrobial (arylamine-containing) sulfonamide will tolerate non-antimicrobial (non-arylamine) sulfonamides [13,23,24] .

The following studies support this statement: A prospective observational study of 94 hospitalized adults with reported "sulfa allergy" noted that 40 patients had taken a non-antimicrobial sulfonamide as outpatients for a median duration of 6.2 years, and nine patients had received non-antimicrobial sulfonamides as inpatients, all without adverse reactions [23] . In the majority of these cases, furosemide was the specific non-antimicrobial sulfonamide that was tolerated. A retrospective series described 34 patients with reported "sulfa allergy" who were treated with furosemide and/or acetazolamide for intracranial hypertension [24] . There were no reactions in those who received furosemide. Urticaria developed in two patients treated with acetazolamide, although no patients experienced severe reactions.

However, because non-antimicrobial sulfonamides can occasionally cause hypersensitivity reactions, it is not possible to state definitively that there is no cross-reactivity between these groups of sulfonamides.

Celecoxib — Cross-reactivity between antimicrobial sulfonamides and celecoxib (a non-arylamine sulfonamide) has not been adequately analyzed. However, in keeping with the general lack of cross-reactivity between the antimicrobial and non-antimicrobial sulfonamides, the bulk of evidence to date suggests that cross-reactivity is unlikely [25-27] . As examples: A prospective pilot study of 28 patients with a history of reactions to antimicrobial sulfonamides and indications for treatment with celecoxib were all challenged with this medication, initially with a small test dose and followed by a full dose, after extensive allergy evaluation [26] . All patients initially tolerated the celecoxib uneventfully. Later follow-up documented that a total of five patients discontinued this therapy for a variety of reasons, although none were allergic in nature. In a large meta-analysis of over 11,000 patients from 14 double-blinded trials of celecoxib used for arthritis, the rate of allergic reactions specifically to celecoxib was not statistically different from placebo or active alternate therapy (other non-steroidal anti-inflammatory drugs, NSAIDs) [27] . Although sulfonamide hypersensitivity was part of the exclusion criteria for those trials, 135 patients were included despite histories of sulfonamide hypersensitivity. Subgroup analysis of those patients did reveal a higher rate of dermatologic reactions compared to the group as a whole. However, the 3- to 6-fold elevation in rate of reactions was also seen in patients receiving other NSAIDs and placebo, indicating that these patients were at higher risk for hypersensitivity drug reactions in general [27] .

However, there are separate concerns about celecoxib and an increased incidence of SJS. A large retrospective review of post-marketing safety surveillance reports (from the United States Food and Drug Administration Adverse Events Reporting System) for several of the selective COX-2 inhibitors documented significant increases of reported SJS above baseline rates [28] . Compared with an expected background rate of 1.9 cases per million annually, calculated rates were 5.7 per million in celecoxib users (based on 19 reported cases in the first two years of marketing). Twenty-one percent of these individuals had also reported "sulfa allergy." Since comparison data from appropriate control groups are not available in studies of this nature, risks of possible cross-reactivity cannot be calculated or directly inferred. However, further studies are needed to define the risk of SJS with celecoxib.

In summary, although there are few data to support cross-reactivity between sulfonamide antimicrobials and celecoxib, celecoxib may be associated with an increased rate of SJS. Until more information is available, we would consider the use of celecoxib in patients with past mild sulfonamide hypersensitivity, although we would avoid celecoxib in a patient with a past severe hypersensitivity to sulfonamides or exfoliative reaction to any medication.

Between sulfonamide antimicrobials and dapsone — As previously mentioned, dapsone is a sulfone, not a sulfonamide. Hypersensitivity to dapsone (or the "sulfone syndrome") is characterized by symptoms similar to those seen in sulfonamide reactions (such as fever and rash, sometimes accompanied by hepatitis, lymphadenopathy, and/or hemolytic anemia) and can be even more severe [7,29,30] . It is currently unclear if there is cross-reactivity between sulfones and sulfonamide antimicrobials.

There is a high rate of dapsone intolerance among HIV-infected patients, although it is not as high as the rate of reactions to TMP-SMX in most studies [31,32] . The ability of HIV infected patients with a history of hypersensitivity to TMP-SMX to tolerate dapsone has been evaluated [31,33-35] : Several studies documented that a significant number of HIV-infected patients who reacted to TMP-SMX also reacted to dapsone, such as in 5/75 (7 percent), 9/23 (39 percent), and 13/60 (22 percent) [31,33-35] . In a direct cross-over trial, 86 patients with HIV who required PCP prophylaxis were randomly assigned to dapsone or TMP-SMX, and then "crossed over" to the other therapy if they reacted to the first agent [35] . Among those who reacted to the initial agent, only approximately one-half of each group was able to tolerate the other medication, for a cross-reactivity rate of about 50 percent. Intolerance was not specifically limited to presumed allergic mechanisms, however.

Among non-antimicrobial sulfonamides — The non-antimicrobial sulfonamides have been associated with far fewer immunologic reactions. Drug eruptions and anaphylaxis have been reported, although not convincingly documented [36,37] . Information about cross reactivity among various agents within this class is limited to case reports [36,37] . It can be concluded from the paucity of reports that cross-sensitivity is very low. Thus, when justified by clinical need, a patient with a past non-anaphylactic reaction to a non-antimicrobial sulfonamide can receive a non-antimicrobial sulfonamide from another group using a graded challenge protocol. (See "Sulfonamide diuretic allergy in a patient needing diuresis" below).

EVALUATION AND DIAGNOSIS OF 'SULFA' ALLERGY — A thorough and detailed history is the most important component of the evaluation of a patient who reports a "sulfa" allergy because there are no definitive tests for either confirming or excluding true sulfonamide allergy.

History — When obtaining the history, focused and specific questions should be directed to address all relevant prior adverse drug events [38] . Details should include: Each exact agent reportedly associated with a reaction, and the indication for its use The precise nature and severity of that reaction The dose and route of medication taken (if known) Other concurrent medications and their prior duration of use The timing of onset of the reaction (from both the precipitating dose, as well as from the initiation of that course of therapy) Any treatment given and response to that treatment (including the duration of reaction) Length of time elapsed since last known exposure to medications of that same pharmaco-immunologic class Any prior or subsequent history of exposure to agents in that class, and presence or absence of any associated reaction(s)

Sometimes, obtaining further details from relatives may be particularly helpful, especially if the patient was an infant or child at the time. Any available medical records of the reaction are invaluable for information about vital signs, objective skin findings, and other important historical data.

The primary purpose of the history is to assess what type of reaction the patient likely experienced in the past, and to make sure that the reaction was not suggestive of severe forms of sulfonamide-induced allergy, such as SJS, TEN, or anaphylaxis. Accordingly, patients who report emergency department visits or hospitalizations, blistering or peeling of the skin or mucous membranes, or use specific phrases such as "anaphylaxis" or "nearly died," in association with prior drug reactions, must be taken very seriously [2] .

Referral — Because of the potentially life-threatening severity of some reactions to sulfonamide drugs, consultation with an allergy specialist experienced in adverse drug reactions is recommended if any form of re-exposure or the use of another sulfonamide or sulfone is under consideration. The management of mild reactions to TMP-SMX in patients with HIV infection may be an exception to this recommendation, as the clinicians treating these patients are often well-versed with these reactions and their management. However, if a patient has experienced a significant or multi-system reaction to a sulfonamide, cannot provide much information, or does not have co-existent HIV infection, then we strongly recommend referral to an allergy specialist if retreatment is being considered.

Testing — There are currently no reliable, valid, and practical testing techniques to evaluate patients with past sulfonamide reactions. A variety of immunologic testing techniques have been tried experimentally, including in vivo tests such as skin testing and patch testing, as well as in vitro tests such as lymphocyte toxicity assays, ELISA, RAST, and specific IgG and IgE testing [39-44] .

Consideration of trimethoprim allergy — A subset of patients who react to a TMP-SMX combination product are sensitive to the trimethoprim component. In two studies of patients with HIV and history of reactions to TMP-SMX, 5 of 31 (16 percent) and 14 of 73 (19 percent) reacted to TMP alone upon oral rechallenge [45,46] . Challenge with trimethoprim to clarify the situation can be helpful, as patients with trimethoprim allergy should be spared the ramifications of being labeled allergic to sulfonamides [46] .

Diagnosis — The diagnosis of the most common form of sulfonamide allergy is made clinically based upon the characteristic pattern of subacute onset of a compatible rash and fever, occurring after exposure to a potential offending antimicrobial agent [47] . (See "Fever and morbilliform rash" above).

SJS and TEN are also clinical diagnoses supported by compatible histologic findings. (See "Stevens-Johnson syndrome and toxic epidermal necrolysis: Clinical manifestations, pathogenesis, and diagnosis", section on diagnosis).

Issues surrounding the diagnosis of other types of allergic reactions to sulfonamides are discussed separately. (See "Serum sickness and serum sickness-like reactions" and other topic reviews).

PREVENTION OF RECURRENT REACTIONS — There are three general approaches to the prevention of recurrent reactions in patients with a possible drug allergy: use of a different class of medication, test dosing, and desensitization [48] . The choice of strategy depends upon the type and severity of the reaction the patient experienced in the past, and on the risks and benefits of alternative treatments. (See "Options in patients with previous morbilliform rash and fever" below and see "Options in patients with previous SJS or TEN" below).

General strategies Use of a non-sulfonamide, non-sulfone medication — Use of an alternative medication of a different (and immunologically non cross-reactive) pharmacologic class is the safest and most suitable option in most patients with past sulfonamide reactions. Safe and effective alternatives are commonly available for most situations, except for HIV-infected patients requiring prophylaxis for Pneumocystis carinii (P. jirovecii). Test dosing — Test dosing (or graded drug challenges) is a term used to describe the cautious administration of small and increasing doses of a medication under careful medical observation. This is an appropriate conservative approach for patients with clinical histories that are unconvincing for true immunologically-mediated drug allergy, or in patients whose reactions were self-limited and remote in time. Test dosing is also useful in patients who require a medication that is believed to be only weakly or uncommonly cross-reactive with the drug that caused the original reaction.

General guidelines for starting doses, dose escalation, frequency of dosing, and similar details have been published, although these are dependent upon the specifics of each case and require individualized modification [4,48] . The starting dose used is influenced by the severity and remoteness of the past reaction, and may range from 1/1000 to 1/10 of the therapeutic dose. This amount is then gradually increased by 2- to 10-fold at each step. The steps are administered at time intervals that are long enough to allow for the development of symptoms, and these intervals are determined by the nature of the past reaction. Desensitization/hyposensitization — The term desensitization is used to describe the cautious administration of small and increasing doses of a medication under careful medical observation to patients in whom allergic reactions are considered PROBABLE, as when immunologic tests are positive. Desensitization may be attempted for some types of reactions when there are no medically acceptable alternatives.

The term desensitization usually refers to procedures designed to allow patients with a history of Type I (IgE-mediated, immediate type hypersensitivity) drug allergy to temporarily tolerate readministration of a medication, such as penicillin [11] . Desensitization implies semi-permanence, although this is NOT the case. Patients must continue to take the medication regularly (usually at least daily) following a successful desensitization procedure. For this and other reasons, the term "hyposensitization" is preferred by some allergy experts.

Protocols have also been developed for a small number of other medication reactions that result from non-IgE mediated immunologic mechanisms, including sulfonamide sensitivity involving delayed maculopapular rashes and fever. Unfortunately, these procedures for non-IgE mediated skin rashes are also called desensitizations, despite the lack of evidence that any immune cell (eg, T lymphocyte) is downregulated as a result of the process. Despite these uncertainties, desensitization is the commonly-used term for sulfonamide re-exposure protocols, and this review uses the term as well.

These protocols may appear to be successful in large part because many patients lose immunologic sensitivity over time. The immune reactivity of HIV-infected patients, in particular, can vary over time. However, "desensitization" has been shown to be superior to simple reintroduction, even in patients with relatively mild reactions. A randomized trial compared a six-step dose escalation protocol with immediate resumption of one TMP-SMX SS tablet daily in 190 HIV-infected patients who had previously discontinued TMP-SMX for non life-threatening rash and/or fever [49] . All patients received concomitant antihistamines. The primary endpoint was the ability of the subject to tolerate the medication daily for six months. The trial was stopped early because of a higher rate of success in the dose escalation group (75 versus 57 percent). Thus, use of a desensitization protocol is warranted. This study also demonstrated that approximately 50 percent of patients with HIV infection and past rash and/or fever would be able to tolerate simple readministration. Note that patients with desquamative rash/SJS/TEN or anaphylaxis were not included in this study.

Precautions with any form of re-exposure — Consultation with an allergy specialist experienced in adverse drug reactions is recommended if either test dosing or desensitization are under consideration, unless the clinician is specifically experienced with these procedures. Precautions regarding both test dosing and desensitization include the following: These procedures are performed immediately before required treatment. Medications are administered without premedication, which could obscure detection of early and mild allergic symptoms. It is essential that any other concurrent health issues are under optimal control, particularly cardiopulmonary conditions. Patients should optimally not be taking medications that may increase the likelihood of anaphylaxis or interfere with treatment of such reactions, such as ACE-inhibitors or beta-blockers. These interventions should be done in an appropriate medical setting, with proper monitoring in place, and immediate availability of rescue medications and equipment. Documentation of informed consent, including a thorough discussion of risks and benefits of the procedure, is essential.

OPTIONS IN PATIENTS WITH PREVIOUS MORBILLIFORM RASH AND FEVER — In most patients with fever and non-blistering rash in response to a sulfonamide drug in the past, the use of an alternative non-sulfonamide, non-sulfone, medication is the most prudent option.

However, if there is no equivalent alternative medication and another sulfonamide medication is required, then both test dosing and desensitization may be considered.

Requiring a sulfonamide from a different group — The best approach in patients who previously reacted to an antimicrobial sulfonamide and require a non-antimicrobial sulfonamide (or reacted to a non-antimicrobial and require an antimicrobial sulfonamide), has not been demonstrated in controlled trials.

We suggest that non-antimicrobial sulfonamides can be administered safely to a patient with a past reaction to a sulfonamide antibiotic (or visa versa) because there is minimal evidence for cross-reactivity between these two groups of sulfonamides. It is our practice to administer one or more test doses of the desired drug first when possible, since rare hypersensitivity reactions to the non-antimicrobial sulfonamides have been reported, and the lack of reported cross-reactivity between these groups of drugs is not proof that it cannot occur. However, if a patient's past reaction to a sulfonamide antibiotic was very mild and remote, test dosing may not be necessary. Similarly, in situations where test dosing is impractical because therapeutic doses of the medication are required urgently (such as emergent diuresis of a patient with past hypersensitivity to TMP-SMX), it may be necessary to administer the sulfonamide in question normally, without test dosing. As mentioned previously, these situations require clinical judgment and clear communication with the patient.

The specific test dose protocol would be based upon the patient's past reaction. If the original reaction involved delayed-onset rash, initial test doses can be repeated for several days without dose escalation in order to ensure that any reactions have sufficient time to develop. As an example, a patient with past hypersensitivity to TMP-SMX who required furosemide could take one-quarter of the lowest strength tablet available, once daily for a week, as an initial test dose. If no symptoms developed, then the patient could assume the therapeutic dose. (See "General strategies" above).

Requiring a sulfonamide from the same group — In patients who previously reacted to an antimicrobial sulfonamide when no satisfactory alternative antibiotic is available (ie, PCP prophylaxis in AIDS patients), a desensitization protocol may be attempted. Desensitization is best studied in HIV-infected patients who require readministration of TMP-SMX, as reviewed in the next section. However, the authors have performed TMP-SMX desensitizations in limited numbers of non HIV-infected patients with similar success.

In patients who developed hypersensitivity reactions to a non-antimicrobial sulfonamide in the past, we suggest avoiding other non-antimicrobial sulfonamides, as the potential for cross reactivity is poorly characterized and newer or alternative therapies are available in most cases.

TMP-SMX allergy in HIV infection — In HIV-positive patients who reacted to TMP-SMX in the past, an alternative agent should be considered initially. Alternative agents include dapsone (although this may also be problematic in patients with sulfonamide hypersensitivity, as previously discussed), inhaled pentamidine, and atovaquone. These and other treatment options are presented in detail elsewhere. (See "Prophylaxis against Pneumocystis carinii (P. jirovecii) in HIV-infected patients").

If alternative agents fail or are not considered equivalent therapeutic options, then desensitization to TMP-SMX may be attempted. Protocols are presented below. (See "Trimethoprim-sulfamethoxazole" below).

Sulfonamide diuretic allergy in a patient needing diuresis — Most of the potent loop diuretics (furosemide, bumetanide, torsemide) and other diuretics (including thiazides, and carbonic anhydrase inhibitors, such as acetazolamide) are sulfonamides. These drugs may cause allergic reactions, as discussed previously, albeit less commonly than the antimicrobial sulfonamides [50,51] . The risk of cross-reactivity among them is poorly understood, although there has been a presumption of cross-reactivity among the thiazide diuretics, supported by case reports.

Presently, ethacrynic acid is the only available loop diuretic with no known potential to cross-react with sulfonamide-containing diuretics. It is therefore the preferred agent for an individual who previously reacted to a sulfonamide-containing diuretic and requires diuresis again.

There have also been published reports of successful graded-dose challenges to torsemide and desensitizations to furosemide in patients with previous reactions to sulfonamide diuretics [50-52] . These were useful during a past period of ethacrynic acid unavailability and are options in those who are unable to be effectively diuresed with ethacrynic acid.

Sulfasalazine allergy in inflammatory diseases — Sulfasalazine is a sulfonamide pro-drug that is converted into 5-ASA following administration. It is used in the treatment of inflammatory bowel disease, psoriasis, lupus, and rheumatoid arthritis. A history of sulfonamide hypersensitivity reactions frequently complicated the management of these conditions in the past, although newer preparations of 5-ASA, which are not sulfonamides (eg, mesalamine), and other immunomodulatory therapies have become available [53] . However, prior to these new therapies, relatively successful protocols for sulfasalazine reintroduction were published [54-57] . (See "Medical management of ulcerative colitis" and other appropriate topic reviews).

Sulfonylurea allergy in diabetes — Some of the oral hypoglycemic agents used in Type II ("adult-onset, non-insulin-dependent") diabetes are non-antimicrobial sulfonamides, which are chemically classified as sulfonylureas. These medications occasionally cause adverse reactions with presumed immunologic mechanisms, most commonly dermatologic reactions such as urticaria, photosensitive rash, and leukocytoclastic vasculitis. There is little in the published medical literature addressing issues of possible cross-reactivity between sulfonylureas and other sulfonamides. There may, however, be some distinct patterns of cross-allergenicity within the sulfonylurea group itself [37] .

The issue of sulfonylurea allergy in diabetics is now less of a concern, as numerous other agents of different classes and mechanisms have been developed. New treatment paradigms have since shifted these medications away from their previously central role, given their less favorable overall risk-to-benefit ratio. Achieving glycemic control using parenteral insulin presents another effective management alternative. (See "Initial management of blood glucose in type 2 diabetes mellitus" and other appropriate topic reviews).

PROTOCOLS FOR DESENSITIZATION — The protocols described in this section are intended to be performed in patients with past morbilliform rash and fever, for whom there is no equivalent alternative treatment.

Drug desensitization protocols vary widely, although all are based upon administering increasing doses of medication in a stepwise manner, such that exposure begins at very low doses, increases gradually, and is uninterrupted. It is not clear whether these protocols induce a state of clinical tolerance, or just permit safe re-administration. The protocols presented here have been performed successfully by the author, although modifications may be made based on characteristics of the individual case. In most cases, these modifications are made to escalate the dose more slowly. Clinicians should avoid altering successful protocols to render them faster or significantly change the time interval between doses.

Trimethoprim-sulfamethoxazole — There are a significant number of published desensitization protocols for use in HIV-positive patients requiring TMP-SMX, 11 of which were compared in detail in a review on drug desensitization [11] . No studies have compared these protocols in a head-to-head manner and the optimal approach has not been determined. Most of the published TMP-SMX desensitization trials have used the oral route, as this is presumed safer than intravenous administration.

Two prospective trials of TMP-SMX desensitization are presented in detail here [43,58] . Both involved HIV-positive patients who required PCP prophylaxis and had a history of TMP-SMX associated fever or non-exfoliative skin rash. We have used these two protocols successfully in our clinic. In addition, we have performed these desensitization protocols in patients without HIV infection, with no apparent differences in outcome. The largest published study of TMP-SMX desensitization in HIV-positive patients involved 48 patients undergoing the protocol outlined in the table (show table 2) [58] . By day three, subjects were receiving a full (single strength) TMP-SMX tablet, and 77 percent of those undergoing this regimen were able to tolerate the course of therapy for a mean of 16 months without adverse effects. Eleven subjects failed to complete the course of therapy, with eight stopping within the first two days. One patient experienced hypotension and acute nonfatal myocardial infarction. Higher relative and absolute CD4 counts were associated with trends towards higher risk of failure. Another study of 22 patients used a longer nine-day protocol, which was successful in 87 percent of patients (show table 3) [43] . Reactions occurred in eight individuals (36 percent). All of these reactions consisted of mild to moderate rash or fever, which resolved after a brief course of systemic glucocorticoid treatment. There were no episodes of hypotension or anaphylaxis.

We initiate all desensitization protocols under medical supervision. For patients with mild fever and/or morbilliform rash in the past, we typically administer the first three doses of this protocol in the clinic, and then allow the patient to complete the remaining steps as an outpatient. The three-day protocol is sufficiently short that it can be administered entirely in the clinic, under medical supervision. This may be preferred, for example, if the patient's ability to follow a timed schedule of doses is uncertain. In contrast, the longer nine-day protocol may be better for reliable patients in whom the main concern is introducing the medication very gradually, as the success rate of this protocol was high.

Symptoms during desensitization — Any symptoms that develop during desensitization must be promptly evaluated in the clinic. As an example, if a patient develops mild morbilliform rash during desensitization, without fever, systemic symptoms, or mucosal involvement, then we typically reduce the dose to the last tolerated step, and administer this dose at the usual time intervals, while simultaneously treating the rash. We treat mild rashes with antihistamines and more extensive rashes with oral glucocorticoids. When the rash begins to improve, the doses can be escalated again, although we sometimes increase the doses more gradually. We typically abort the desensitization protocol if fever or systemic symptoms develop. More rapid protocols have also been published for TMP-SMX [59] . These may warrant additional caution for potentially increased risks.

One study suggested that desensitization may not be necessary in some HIV-positive patients with very mild past reactions to TMP-SMX. This multi-center, open study compared sulfonamide desensitization to cautious test dosing [46] . Reaction rates were slightly lower in the desensitization group (21 versus 28 percent), although the difference was not statistically significant. This suggests that some patients who had reacted in the past were not actually at risk for recurrent reactions, either due to immunologic changes over time, inaccurate or incorrect clinical histories, or other unknown factors. It is also possible that the study was underpowered to detect a small benefit of desensitization. Furthermore, the findings of such studies must be interpreted with caution, as patients with severe reactions are specifically excluded, and selection bias must be considered. Thus, if TMP-SMX is to be administered to a patient with a past reaction, we perform desensitization, rather than simply reintroducing the drug, even in patients with relatively minor past reactions.

Risks — These desensitizations are generally well tolerated [11] . However, risks of drug desensitization procedures include, but are not limited to, the entire spectrum of allergic reactions. This includes dermatologic reactions such as urticaria/angioedema or other types of rash, pulmonary reactions such as asthma, and rare multi-system anaphylaxis. Other types of serious and life-threatening reactions have also been reported in association with these procedures, including myocardial infarction [58] .

Duration of efficacy — Desensitization may be effective only as long as the allergic individual is continuously exposed to the drug. After completing a desensitization protocol for an IgE-mediated drug reaction, the patient must continue the course of treatment without missing a single daily dose [60] . Whether the same is true for patients desensitized to TMP-SMX, however, is speculative. Patients should be counseled that they may have a recurrent reaction if they forget to take their daily dose on one or more consecutive days.

Other sulfonamides — Desensitization protocols for sulfonamides other than SMX, such as sulfadiazine (for cerebral toxoplasmosis in patients with AIDS) or sulfasalazine (for patients with rheumatoid arthritis), have been published as well [57,59] . (See "Sulfasalazine in the treatment of rheumatoid arthritis").

Dapsone — Successful "desensitization" to dapsone has been reported in HIV-positive patients [61] . One case report described successful dapsone desensitization in a patient who reacted to both TMP-SMX and dapsone, and failed TMP-SMX desensitization [62] .

OPTIONS IN PATIENTS WITH PREVIOUS SJS OR TEN — As discussed previously, exfoliative rashes include SJS and TEN. (See "Stevens-Johnson syndrome and toxic epidermal necrolysis: Clinical manifestations, pathogenesis, and diagnosis").

ANY form of re-exposure to the culprit drug or a sulfonamide in the same group is strictly contraindicated if a patient describes symptoms consistent with an exfoliative dermatitis or diffuse erythroderma. This contraindication applies to both test dosing and desensitization. Case series have noted that SJS and TEN can recur with even minor re-exposures and may be more severe the second time [63] .

Requiring a sulfonamide from a different group — There are no studies that provide guidance in the management of patients who experienced SJS/TEN to a sulfonamide medication in the past and truly require another sulfonamide medication. Our practice is to avoid all sulfonamides and sulfones in such individuals. (See "Stevens-Johnson syndrome and toxic epidermal necrolysis: Management, prognosis, and long-term sequelae").

However, there may be rare situations in which the use of a sulfonamide from a different group is needed for a life-threatening condition. As an example, it is conceivable that a patient with SJS to TMP-SMX in the past may require urgent diuresis with a potent loop diuretic (such as furosemide) after failing to diurese adequately with the non-sulfonamide drug, ethacrynic acid. Consultation with an allergy specialist experienced in adverse drug reactions would be essential in this setting.

INFORMATION FOR PATIENTS — Educational materials on this topic are available for patients. (See "Patient information: Allergy to penicillin and related antibiotics"). We encourage you to print or e-mail this topic, or to refer patients to our public web site www.uptodate.com/patients, which includes this and other topics.

SUMMARY AND RECOMMENDATIONS

Clinical characteristics and pathophysiology There are two distinct groups of sulfonamides: antimicrobial sulfonamides and non-antimicrobial sulfonamides (show figure 2A-2B). Antimicrobial sulfonamides contain an arylamine group that undergoes metabolic changes believed critical to the development of hypersensitivity reactions. Medications in the non-antimicrobial group do not contain arylamine groups and are associated only rarely with hypersensitivity reactions. Dapsone is not a sulfonamide, although it can cause severe hypersensitivity reactions (show figure 3). (See "Terminology" above). Sulfonamide reactions are most prevalent in individuals with untreated HIV disease, and especially in those with active AIDS. Reaction rates appear to be declining since the introduction of HAART. (See "Epidemiology" above). Hypersensitivity reactions to sulfonamide antimicrobials most commonly present as delayed-onset fever and morbilliform rash, sometimes accompanied by organ involvement. (See "Types and clinical manifestations of hypersensitivity reactions" above). Sulfonamides less commonly cause other types of immunologic drug reactions, including serum sickness-like reactions and IgE-mediated allergic reactions. Sulfonamide antimicrobials are the most frequently implicated agents in the rare reactions, Stevens-Johnson Syndrome (SJS) and toxic epidermal necrolysis (TEN). (See "Types and clinical manifestations of hypersensitivity reactions" above). The pathophysiology of hypersensitivity reactions to sulfonamides is incompletely understood. Metabolism of sulfonamide antimicrobials at the arylamine group results in intermediate forms of the drug that are immunologically reactive and/or directly cytotoxic. Factors that slow metabolism, including slow acetylation and/or glutathione deficiency states (eg, in HIV infection) may increase exposure to these metabolites and thus enhance the likelihood of hypersensitivity reactions (show figure 4). (See "Pathophysiology" above).

Issues of cross-reactivity There is minimal evidence of cross reactivity between the antimicrobial sulfonamides and the non-antimicrobial sulfonamides (show figure 2A-2B). However, it is impossible to say with certainty that cross-reactivity does not occur, because both groups have also been implicated in serious hypersensitivity reactions, and because drug reactions disproportionately affect certain patients who react to multiple unrelated drugs (ie, "multiple drug allergy syndrome"). (See "Between sulfonamide antimicrobials and non-antimicrobials" above). There is minimal evidence for cross-reactivity between antimicrobial sulfonamides and celecoxib, and patients who react to both appear to be rare. However, celecoxib may be independently associated with a higher incidence of SJS. (See "Celecoxib" above). In HIV-infected patients, intolerance to both TMX-SMX and dapsone is common, although there is no clear evidence of cross-reactivity. (See "Between sulfonamide antimicrobials and dapsone" above). There is only anecdotal evidence of cross reactivity among different non-antimicrobial sulfonamides, and it is not well characterized. One important exception to this statement is the potential cross reactivity between different sulfonamide-containing diuretics, which is suggested by case reports. (See "Among non-antimicrobial sulfonamides" above).

Evaluation and diagnosis — The evaluation and diagnosis of a patient with a past sulfonamide reaction is based solely upon a detailed clinical history, because there are no useful in vitro or in vivo tests. (See "Evaluation and diagnosis of 'sulfa' allergy" above).

Management of patients with past reactions — In patients with a past reaction to a sulfonamide, the use of an alternative medication is the most suitable option in most cases, although the potential consequences of treatment failure with a "second-best" therapy must also be considered. (See "Prevention of recurrent reactions" above). An allergy specialist should be consulted (if possible) when any form of re-administration of the culprit agent or a similar medication is being considered, unless the clinician is experienced in these procedures (eg, HIV specialists). (See "Precautions with any form of re-exposure" above).

Fever and non-exfoliative rash — Patients with a history of sulfonamide hypersensitivity (characterized by fever and non-exfoliative morbilliform rash) should avoid the causative agent and other agents in the same group (eg, within the antimicrobial arylamine group) when acceptable alternatives are available. (See "Options in patients with previous morbilliform rash and fever" above). If the culprit agent or another agent within the same group is clinically required, we suggest performing a desensitization protocol (rather than simple readministration or test dosing) (Grade 2B). Each case requires careful consideration of risks versus benefits, particularly if the original reaction involved extracutaneous organ inflammation. (See "Requiring a sulfonamide from the same group" above). Desensitization has been best studied in HIV-positive patients who require readministration of TMP/SMX, although the optimal desensitization protocol for this situation has not been determined. We typically use one of two protocols in our clinic (show table 2 and show table 3). (See "Protocols for desensitization" above). Desensitization protocols have been published for a small number of non-antimicrobial sulfonamides, including torsemide, furosemide, and sulfasalazine, as well as for dapsone. (See "Other sulfonamides" above and see "Dapsone" above). If the patient requires a sulfonamide medication from a different group, then we suggest administering it (Grade 2C). We administer one or more test doses initially, rather than simply giving the medication normally, although the necessity of this extra step is not known. (See "Requiring a sulfonamide from a different group" aboveSee "Requiring a sulfonamide from a different group" above).

SJS/TEN — Patients with a history consistent with SJS, TEN, or an exfoliative dermatitis due to a sulfonamide medication should strictly avoid the culprit drug and other agents in the same sulfonamide group. Re-exposure to the same agent may be fatal. We suggest such patients avoid all other sulfonamides and sulfones whenever possible (Grade 2C). (See "Options in patients with previous SJS or TEN" above).

ACKNOWLEDGMENT — The author and editorial staff at UpToDate, Inc. would like to acknowledge Donald Dibbern, MD, for his contributions to an earlier version of this topic review.